Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/10628738

TheMolecularBasisofVitaminERetention:StructureofHumanα-TocopherolTransferProtein

ArticleinJournalofMolecularBiology·September2003

DOI:10.1016/S0022-2836(03)00724-1·Source:PubMed

CITATIONS

110

READS

90

5authors,including:

Someoftheauthorsofthispublicationarealsoworkingontheserelatedprojects:

ReactionMechanismofChromanol-RingFormationViewproject

PropertiesofF1-ATPasecomplexesViewproject

TakashiTomizaki

57PUBLICATIONS5,132CITATIONS

SEEPROFILE

ClemensSchulze-Briese

DECTRISLTd.

107PUBLICATIONS2,198CITATIONS

SEEPROFILE

AchimStocker

UniversitätBern

73PUBLICATIONS1,973CITATIONS

SEEPROFILE

AllcontentfollowingthispagewasuploadedbyAchimStockeron10November2017.

Theuserhasrequestedenhancementofthedownloadedfile.

The Molecular Basis of Vitamin E Retention: Structureof Human a-Tocopherol Transfer Protein

Reto Meier1, Takashi Tomizaki2, Clemens Schulze-Briese2

Ulrich Baumann1 and Achim Stocker1*

1Department of Chemistry andBiochemistry, University ofBerne, Freiestrasse 3, 3012Berne, Switzerland

2Paul-Scherrer-InstitutCH-5232 Villigen PSISwitzerland

a-Tocopherol transfer protein (a-TTP) is a liver protein responsible for theselective retention of a-tocopherol from dietary vitamin E, which is amixture of a, b, g, and d-tocopherols and the corresponding tocotrienols.The a-TTP-mediated transfer of a-tocopherol into nascent VLDL is themajor determinant of plasma a-tocopherol levels in humans. Mutationsin the a-TTP gene have been detected in patients suffering from lowplasma a-tocopherol and ataxia with isolated vitamin E deficiency(AVED).

The crystal structure of a-TTP reveals two conformations. In its closedtocopherol-charged form, a mobile helical surface segment seals thehydrophobic binding pocket. In the presence of detergents, an open con-formation is observed, which probably represents the membrane-boundform. The selectivity of a-TTP for RRR-a-tocopherol is explained fromthe van der Waals contacts occurring in the lipid-binding pocket.Mapping the known mutations leading to AVED onto the crystal structureshows that no mutations occur directly in the binding pocket.

q 2003 Elsevier Ltd. All rights reserved

Keywords: ataxia; AVED; crystal structure; tocopherol; vitamin E*Corresponding author

Introduction

Vitamin E is a general term for a group of lipo-philic compounds, referred to as peroxyl-radicalscavengers and chain-breaking antioxidants withinbiological membranes.1 Eight different vitamin Eforms occur in nature: a, b, g, and d-tocopherolhaving a phytyl tail with three chiral centers inR-configuration and a, b, g, and d-tocotrienolhaving an isoprenoid side-chain.2 Vitamin Edeficiency leads to severe degenerative diseasessuch as ataxia, infertility and Duchene-like muscledegeneration. The connection between these patho-logical processes and vitamin E is not known indetail but oxidative stress is likely to play a majorrole.

Absorption studies using vitamin E isotopeshave led to the discovery of a remarkable prefer-ence for natural RRR-a-tocopherol (RRR-a-T) inhealthy adults.3 The isoforms and stereoisomers ofvitamin E are taken up in equal amounts in an

unspecific process including emulsificationtogether with food-derived lipids and the sub-sequent absorption of the formed lipid particles.4

From the intestine they are transported in chylo-microns and subsequently in remnants to the liverwhere the cytosolic protein a-TTP is responsiblefor the stereo-selective transfer of RRR-a-T toVLDL, which is then released into the circulation.5

In vitro as well as in vivo assays have shown thata-TTP binds preferentially to the RRR-a-Tisomer.6,7

Sequence analysis8 classifies a-TTP as a memberof the widespread SEC14-like protein family har-boring a characteristic CRAL_TRIO lipid-bindingdomain.9 Other members of this family includephosphatidylinositol/phosphatidylcholine transferprotein (SEC14) from Saccharomyces cerevisiae,10 cel-lular retinaldehyde binding protein (CRALBP)11

and supernatant protein factor (SPF).12 All theseproteins appear to mediate the intracellular distri-bution of specific lipids through the aqueousenvironment by a common mechanism.13

The prominent role of a-TTP as carrier of food-derived RRR-a-T has meanwhile been documentedin numerous studies.14 Its importance in maintain-ing normal plasma a-tocopherol concentrationshas been confirmed by analyzing mutations in the

0022-2836/$ - see front matter q 2003 Elsevier Ltd. All rights reserved

E-mail address of the corresponding author:achim.stocker@ibc.unibe.ch

Abbreviations used: a-TTP, a-tocopherol transferprotein; a-T, a-tocopherol; AVED, ataxia with isolatedvitamin E deficiency.

doi:10.1016/S0022-2836(03)00724-1 J. Mol. Biol. (2003) 331, 725–734

gene of a-TTP in patients suffering from ataxiawith vitamin E deficiency (AVED).15

Besides high expression rates of a-TTP in theliver, the mammalian brain seems to be able toexpress its own a-TTP. Accordingly, the neurologi-cal phenotype of a-TTP2/2 mice has been foundto be even more severe and shows an earlier onsetthan that of wild-type mice when maintained on aa-T-deficient diet,16 the severe phenotype beingunable to walk straight forward. Moreover, auterine form of a-TTP has been reported to beessential for embryogenesis by supplying thelabyrinth region of the placenta with RRR-a-Tduring development.17 Both tissues are known tobe exposed to high rates of oxidative stress andtherefore seem to be specifically protected bya-TTP-mediated tocopherol delivery.

Here we present the crystal structures of a closed“carrier” conformation and an open “membrane-docking” conformation of human a-TTP bound toits physiological ligand RRR-a-T and to the deter-gent Triton X-100, respectively. This analysisprovides the molecular basis of vitamin E reten-tion. Furthermore, it elaborates a model of lipidtransfer and opens a way for the rationalization ofmutations leading to AVED.

Results

Overall structure of a-TTP

a-TTP crystallizes in two different space groupscorresponding to distinct conformational states ofthe molecule. In the presence of Triton X-100,crystals belonging to the monoclinic space groupP21 were obtained with two a-TTP molecules perasymmetric unit. Addition of RRR-a-T con-comitant with removal of Triton X-100 resulted ina tetragonal crystal form containing one moleculeper asymmetric unit. The overall conformation ofa-TTP is quite similar in both crystal forms withthe exception of the helical segment 198–221(Figure 1).

Our current a-TTP model comprises residues9–273 in the monoclinic and residues 25–275 inthe tetragonal crystal form (Figure 2). In the latter,the N-terminal helix would interfere with a crystalcontact and is completely disordered. The proteinfolds into the prototypical SEC14-like domainstructure, which encompasses two highly con-served structural elements, an amino-terminalthree-helix coil and a C-terminal CRAL_TRIOlipid-binding domain.18,19

Figure 1. Overview of the open (left) and of the closed (right) conformation of the a-TTP structure. The left picturecorresponds to the non-crystallographic dimer. The a-TTP domain backbone is shown in a cartoon representation,with helices colored in blue, b-strands in orange and the lid colored in gray. Triton molecules (left) and RRR-a-toco-pherol (right) are shown as stick models with carbons colored yellow. This Figure was prepared using the programPYMOL (Warren Delano, http://www.pymol.org).

726 Structure of Human a-Tocopherol Transfer Protein

Functional assignment of the open and theclosed conformation of a-TTP

The greatest difference between the two crystalforms occurs in the solvent-exposed residues 198–221, termed “lid”. The conformation of this lidaffects ligand access to the lipid-binding site. Inthe RRR-a-T complex the ligand is deeply buriedin a hydrophobic pocket, which is closed by thelid. The hydrophobic side of the lid helix lies onthe entrance of the pocket while the more polarone faces the solvent (Figure 3).

On the other hand, the Triton complex of a-TTPadopts a conformation with an open lid. Thehydrophobic side (residues Ile202, Phe203, Val206and Ile210) of the lid helix now faces the solvent.This hydrophobic patch is partially covered by aTriton X-100 molecule. A further Triton X-100molecule is bound within the hydrophobic pocketwhere it replaces RRR-a-T. The two lids of theNCS-related molecules interact with each otherforming the dimer contact area. Interestingly, asimilar contact is found in the crystal lattice of theSEC14 structure,20 which crystallizes in the

Figure 2. Stereoview of labeled Ca traces of the open (blue and cyan) and of the closed (red and orange)conformation of the a-TTP structure.

Figure 3. Surface comparison of the open (left) and of the closed (right) conformation of the a-TTP structure. Hydro-phobic residues (Ala, Val, Cys, Leu, Ile, Phe, Tyr, Trp, Pro) are colored yellow, basic residues (Arg and Lys) are coloredblue, acidic amino acids (Glu and Asp) are red, polar residues (Asn, Gln, Ser, Thr, Gly) are shown in cyan. The hydro-phobic key residues of the mobile lipid-exchange loop (residues 198–221) are shown in light gray.

Structure of Human a-Tocopherol Transfer Protein 727

presence of b-octylglucoside in an open confor-mation as well. The rotation of the lid by about808 against the plane of the beta-sheet causes ashift of about 14 A. This large conformationalchange at the entrance of the cavity contrasts withthe absence of significant changes in the rest of themolecule.

The RRR-a-T binding site

The ligand-binding pocket of a-TTP is mostlylined with hydrophobic amino acid side-chainsdefining a cavity (program VOIDOO)21 (Figure 4).The chromanol moiety of RRR-a-T is mostly sur-rounded by hydrophobic residues with the excep-tion of Ser136, Ser140 and three water molecules.One of these connects the tocopherol phenolichydroxyl group in para with the backbone carbonylgroup of Val182 through a hydrogen bond. Severalwater molecules are lined up in a hydrophilictunnel giving access to the bulk solvent.

The aromatic methyl group in the 5 position ofthe chromanol moiety fits snugly into a nicheformed by the side-chains of residues Ile194,Val191, Ile154 and Leu183, the latter two being invan der Waals contact, with a distance of about3.6 A. On the other side of the chromanol ring thetwo aromatic methyl groups in the 7 and 8 positionmake contacts to Phe187, Phe133 and Leu137.

The position and geometry of the pyran half-chair of the chromanol ring determine the relativepositions of the substituents at the stereocenter inthe 2 position with the axial methyl group pro-truding into an indent of the cavity formed by resi-dues Phe133, Val182 and Ile179 (Figure 4B). Theprenyl side-chain is bent into a U-turn involvingboth stereocenters at the 40 and at the 80 position.

Discussion

The structural basis of RRR-a-T selectivity

In vitro tocopherol-transfer studies comparingvitamin E derivatives by Hosomi et al. uncovereda preference of a-TTP for RRR-a-T (100%) relativeto RRR-b-T (38%), RRR-g-T (9%) and RRR-d-T(2%).6 Of these four naturally occurring analogs,RRR-b-T is found in negligible amounts in food,while RRR-d-T, RRR-a-T and RRR-g-T areabundant in different ratios in most edible oils.22

RRR-g-T lacks one aromatic methyl group in the 5position and should therefore fit into the cavity aswell (Figure 4). However, the absence of onemethyl group reduces the surface available forhydrophobic interactions and diminishes thepacking density. Studies by Fersht and co-workersderived an average penalty of 1.3(^0.5) kcal/molfor the removal of a single methylene group from

Figure 4. A, Schematic drawing of the interactions between RRR-a-T and the residues within the tocopherol-bindingpocket. Possible hydrogen bonds and short contacts are shown with broken lines, and van der Waals contacts are indi-cated by arcs. This diagram was generated using the program LIGPLOT51. B, Stereo figure of the ligand-binding pocketof a-TTP with electron density of bound RRR-a-tocopherol. Shown is a 2Fo 2 Fc density map contoured at 1.0s abovethe mean. The map was computed before the ligand was included in the model. RRR-a-tocopherol is depicted as astick model in dark gray, internal water molecules as black balls, residues within van der Waals distance to thetocopherol molecule in gray.

728 Structure of Human a-Tocopherol Transfer Protein

the hydrophobic main core of chymotrypsininhibitor 2.23 Taking this number as a rough guidefor the removal of a methyl group from the hydro-phobic cavity of a-TTP, the binding ratioKg–T=Ka–T is computed as 8.3^6.7 at 310 K. Thisestimate fits the experimentally determined tenfoldreduction in RRR-g-T binding;6 of course the erroris large due to the varying extent of hydrophobiccontacts of a methyl group, which is reflected inthe large uncertainty of the figure given by Fershtand co-workers. In the case of RRR-d-T wherean additional methyl group is missing in the 7position the computed ratio of Kd–T=Ka–T equals92, a value that correlates reasonably well with theexperimentally observed 50-fold reduction inbinding,6 considering again the large error.

In vivo and in vitro experiments using eight indi-vidual stereoisomers of a-tocopherol haveshown that a-TTP possesses stereoselectivitytowards R-configuration in the 2 position of thechromanol ring.5,24,25 The synthetic SRR-diastereo-mer of a-T exhibits a ten times lower affinity fora-TTP than its natural RRR-counterpart,6 mostprobably due to a loss of the packing interactionsdescribed above (see Figure 4B). The stereocentersin the 40 and 80 position of the side-chain werereported to be of minor importance for binding,26,27

which is attributable to fewer packing restraintsand phytyl chain flexibility.

The mechanism of tocopherol transfer

Predictions concerning the transfer mechanismhave been based on two main considerations: (i)the transfer protein has to accomplish physicalcontact with the membrane in a reversible way;and (ii) conformational changes at the protein sur-face mediate the ligand exchange reaction betweenthe hydrophobic cavity of the protein and the acylchain environment of the bilayer. These points cannow be addressed by the two conformationsobserved in our crystal structures.

In the closed conformation a large hydrophobicarea (Phe203, Val 206, Phe207, Ile210 and Leu214)of the lid is in direct contact with the side-chain ofRRR-a-T. Opening of the lid shifts these residuestowards the exterior, e.g. to the membrane surfaceor to the dimer interface in the monoclinic crystalform, establishing new hydrophobic contacts withlipids/detergents or hydrophobic surfaces. Boundtocopherol is then released into the membrane or,vice versa, is shuffled into the empty or water-filledbinding pocket by lid closure. Water molecules inthe cavity, if present, are probably squeezed outthrough a tunnel connecting the rear of the pocketwith the hydrophilic solvent. Well-defined watermolecules are visible at the entrance and withinthe tunnel. The closed lid exposes a more polarface to the solvent and the charged carrier–ligandcomplex can leave the membrane.

This mechanism, i.e. the presence of amembrane-bound open conformation of the carrierprotein where a hydrophobic lid is inserted into

the lipid bilayer, followed by transport in a closedconformation of the protein, has been describedfor the non-homologous phosphatidylinositoltransfer protein a.28,29 The presence of detergents/lipids favors the open conformation also in thiscase, as well as for SEC14 as mentioned above.Furthermore, similar conformational changes havebeen observed in lipases, another class of proteinsdealing with lipid–water interfaces. The lidcomprises an a-helical structure in all cases.28,30

The molecular basis of AVED

Several mutations of the human a-TTP genehave been reported to cause vitamin E deficiency,ataxia and retinitis pigmentosa.15,31,32 The AVEDsyndrome is characterized by deficient plasmavitamin E and by a progressive peripheral neuro-pathy with a specific dying back of the largercaliber axons and the sensory neurons, whichfinally results in ataxia.33 The linkage of the AVEDsyndrome to mutations in the a-TTP gene hasbeen confirmed by the production of a-TTP knock-out mice exhibiting extremely low RRR-a-T levelsand the characteristic neurological disorder.16

In humans a striking correlation between thetype of mutation, the degree of tocopheroldepletion and the severity of the disease has beendescribed.34 Mild changes in phenotypes areobserved in the cases of missense mutations(R192H, H101Q) causing substitutions in non- orsemiconserved amino acids. Patients bearing thesemutations seem to produce a partially functionalprotein, which allows them to incorporate RRR-a-T into plasma at reduced rates, leading to a lateonset of the disease.35 Substitutions of highly con-served amino acids, e.g. R59W, E141K and R221W,are associated with the severe, early onset form ofAVED disease. The severe form is also observed incases of protein truncations deriving from frame-shift mutations, mutations affecting intron/exonboundaries and the nonsense mutation R134X.The most abundant truncation mutation has beenfound in 17 unrelated families with severe AVED,where one base-pair is deleted from position 744(744delA).36 The resulting frame-shift leads to thereplacement of the last 30 amino acid residues byan aberrant 14 amino acid residue peptide affectingthe C-terminal helix (residues 252–272), whichextends behind the b-sheet floor of the hydro-phobic pocket (Figure 2). In SEC14, the missensemutation G266D, which occurs in this region,renders the protein thermolabile for phospholipidtransfer activity in vitro and in vivo by destabilizingthe hydrophobic pocket.10

As shown in Figure 5, the missense mutationE141K in a-TTP disrupts the hydrogen bondbetween the highly conserved E141 on helix 9(residues 129–143) and Y73. This helix is a centralbuilding block of the CRAL_TRIO fold formingone wall of the tocopherol-binding cavity. Thepresence of a positive charge and the lack ofa hydrogen bond in the E141K mutation are

Structure of Human a-Tocopherol Transfer Protein 729

sufficient to cause severe AVED, probably bydestabilizing the CRAL_TRIO fold. Interestingly,on the opposite site of helix 9 a second hydrogenbond between T139 and the semi-conserved H101is involved in the missense mutation H101Q(Figure 5). The replacement of histidine byglutamine may not completely abolish the hydro-gen bond with T139 and thus exhibits less severephenotypes than the E141K mutation.37

Another set of AVED-associated mutationsoccurs in a conspicuous cluster of arginineresidues, namely R59W, R221W and R192H. Thesethree arginine residues are located on a surfacepatch that is highly positively charged. Besidesthese three arginine residues there are furtherpositively charged amino acid residues in thespatial neighborhood, namely R57, R68, R151,K155, K190 and K217. A sulfate ion originatingfrom the crystallization buffer in the tetragonalcrystal form is coordinated to K217, R221, K190and R192. These residues surround the exit of thewater-filled tunnel mentioned above, which con-nects the cavity with the bulk solvent. In two ofthe missense mutations, namely R59W andR221W, a highly conserved arginine residue isreplaced by a hydrophobic tryptophan, while inthe case of the R192H mutation the semi-conservedR192 is replaced by a residue of intermediatepolarity. The first two mutations are associatedwith the severe form of AVED, the latter one witha milder pathology. The consequences of thearginine to tryptophan mutations are the dis-appearance of a positive charge and the placementof a hydrophobic group on the surface. In addition,a salt bridge to D185 is lost in the R59W mutation.The placement of a hydrophobic group on thesurface could cause the formation of protein aggre-

gates while the loss of a salt bridge would beexpected to destabilize the protein fold. One couldalso speculate that the positive charge on thesurface aids docking to the negatively chargedheadgroups of the membrane lipids. Interestingly,the K239A missense mutation in SEC14 completelyabolishes phosphatidylinositol transfer activity invitro.38 This residue is conserved in all SEC14homologs and corresponds to R221 in a-TTP.SEC14 possesses a similar cluster of positivelycharged residues on this surface segment, namelyR63, R65, K66, R119 and K239.

Concluding remarks

Our structural data show that the specificity forRRR-a-T is governed by the packing density withinthe lipid-binding pocket and can be semi-quanti-tatively explained. Two different conformations ofa-TTP are observed, which can be assigned to themembrane-bound form and the carrier state of themolecule, opening a detailed view on the lipid-exchange mechanism. This mechanism seems tobe a general feature of proteins dealing with lipids.

Mutations in the a-TTP gene cause proteinmalfunction, low plasma RRR-a-T levels, and sub-sequently the characteristic AVED.16,35 This severeneurological disorder is observed in cases ofprotein truncations deriving from frame-shiftmutations, mutations affecting intron/exonboundaries and missense mutations. Unexpectedly,none of the known missense mutations of thea-TTP gene affects the lipid-binding pocket or thelid. It may thus be assumed that mutations thatare too prohibitive for tocopherol binding arestrongly counter-selected, presumably for viabilityreasons.

Figure 5. AVED associated mutations in a-TTP. The clinically characterized mutations are mapped onto the three-dimensional structure. Residues that are mutated are shown in green color and carry a label, the interacting aminoacid residues described in the text are shown in brown color without label.

730 Structure of Human a-Tocopherol Transfer Protein

Materials and Methods

Protein expression and purification

All analytical grade chemicals were obtained fromSIGMA (Buchs, Switzerland). The N-terminal (His)6-tagged a-TTP expression construct was made by cloningthe PCR product derived from a human cDNA libraryinto the NdeI and XhoI sites of the pET-28a vector(Stratagene) using the primers 50-GGGAATTCCATATGGCAGAGGCGCGATCCCAG-30 and 50-CCGCTCGAGTCATTGAATGCTCTCAGAAATGC-30. Escheri-chia coli BL21(DE3) STAR cells (Invitrogen, Carlsbad)were grown at 25 8C to an A600 nm of 0.5 and theninduced with 1 mM isopropyl-thiogalactopyranoside(IPTG) for 15 hours. Harvested cells were disruptedtwice in a French press. a-TTP was purified on 10 ml ofNi-NTA SUPERFLOW (Qiagen, Basel, Switzerland)according to the manufacturer’s instructions. The (His)6-tag was cleaved by thrombin digestion (Amersham,Dubendorf, Switzerland) at 4 8C overnight. The proteinconcentration was determined by the microbiuretmethod.39 Typical yields were 0.6 mg/l of pure a-TTP.Higher yields of 8 mg/l were obtained by adding 1%(w/v) Triton X-100 to the disrupted cells and stirringthe suspension for 40 minutes at room temperature.After purification on Ni-NTA, the eluate was adjustedto 0.2% (w/v) Triton X-100 and then concentrated byCentriprep-10 (Millipore, Volketswil, Switzerland) to20 mg/ml. The protein was loaded onto a Superdex 200column (Amersham) in GPC buffer (20 mM Tris,

100 mM NaCl, 2 mM EDTA, 2 mM DTT, pH 8.0) contain-ing 0.05% (w/v) Triton X-100. One major peak represent-ing the a-TTP fraction was pooled and concentrated to20 mg/ml and 0.2% (w/v) Triton X-100. The seleno-methionine (SeMet)-labeled a-TTP was prepared bythe method of methionine biosynthesis inhibition40 andpurified as the wild-type.

Preparation of the RRR-a-T complex with a-TTP

The ligand complex of RRR-a-T with a-TTP was pre-pared as described with minor modifications.41 Briefly,purified a-TTP was diluted in 1.5 l of Tris buffer(10 mM, pH 8.5) to a final concentration of 1 mM andtocopherol was added at a 50-fold molar excess from astock solution of 160 mM RRR-a-T and 100 mM TritonX-100. The diluted protein solution was stirred at 37 8Cfor one hour and then loaded at 4 8C onto a columncontaining 10 ml of High Q (Amersham, Dubendorf,Switzerland). After washing with ten volumes theprotein was eluted with 20 ml of Tris–NaCl buffer(10 mM Tris, 1 M NaCl, pH 8.5). The eluted protein wastransferred into the GPC buffer (20 mM Tris, 100 mMNaCl, 2 mM EDTA, 2 mM DTT, pH 8.0) by Centriprep-10 ultrafiltration and concentrated to 12 mg/ml.

Crystallization and data collection

Monoclinic crystals (Table 1) of a-TTP were obtainedat 18 8C by the sitting-drop, vapor-diffusion method.Drops were set up by mixing 1 ml of protein solution

Table 1. Crystallographic statistics

a-TTP native Triton X100 SeMet-a-TTP Triton X100 a-TTP native RRR-a-T

Crystal parameters P21; a ¼ 45.1 A, b ¼ 113.8 A,c ¼ 67.4 A, b ¼ 98.98two molecules/a.s.ua

P21; a ¼ 45.4 A, b ¼ 113.5 A,c ¼ 67.5 A, b ¼ 99.18two molecules/a.s.u

P41212; a ¼ b ¼ 77.7 A,c ¼ 128.2 A.one molecule/a.s.u

Data collection (XDS)Wavelength (A) 0.9791 0.9791 1.1407No. crystals 1 1 1Resolution range (A) (outer shell) 30–1.88 (1.90–1.88) 20–2.30 (2.35–2.30) 20–1.95 (1.98–1.95)No. observations 345,927 346,225b 317,025No. unique reflections 54,206 58,647b 29,375Completeness (%)c 99.3 (93.1) 99.3 (92.5) 99.7 (97.7)Rsym

d 0.051 (0.443) 0.090 (0.271) 0.062 (0.559)I=s ðIÞ 20.2 (3.2) 12.7 (6.3) 22.8 (3.5)FOMeSAD/solvent flattened 0.32/0.57

Refinement (REFMAC)Resolution range (A) 30–1.88 – 20–1.95No. reflections working set 51,495 – 27,963No. reflections test set 2711 – 1472No. atoms total 4924 – 2208

Protein 4289 2052Lipid 150f 31g

Water molecules 294 125Sulfate ions 0 2R=R-freeð%Þ 19.5/22.6 19.0/22.7RMS bonds (A) 0.010 0.012RMS angles (deg.) 1.3 1.3

a Asymmetric unit.b Friedel pairs were treated as different reflections.c Rsym ¼ ShklSjlIðhkl; jÞ2 kIðhklÞll=ðShklSjkIðhklÞlÞ where Iðhkl; jÞ is the jth measurement of the intensity of the unique reflection ðhklÞ

and kIðhklÞl is the mean over all symmetry-related measurements.d The values in parentheses of completeness, Rsym and I=s ðIÞ correspond to the outermost resolution shell.e Figure-of-merit as computed by SOLVE/RESOLVE.f Six Triton X-100 molecules.g One RRR-a-tocopherol molecule.

Structure of Human a-Tocopherol Transfer Protein 731

(18 mg/ml) with 1 ml of reservoir solution (5% (w/v)PEG 6000, 0.1 M Hepes (pH 7.5), 15% (v/v) MPD) andequilibrated against 100 ml reservoir solution at 18 8C.Crystals were observed after ten hours and grew to anaverage size of 0.25 mm £ 0.07 mm £ 0.015 mm. Iso-morphous crystals of the SeMet-labeled protein wereobtained by mixing 1 ml of protein solution (16 mg/ml)with 3 ml of reservoir solution (2.5% (w/v) PEG 2000MME, 0.1 M Hepes (pH 7.5), 10% (v/v) MPD). Crystalsof the RRR-a-T complex of a-TTP were obtained by mix-ing 1 ml of protein–tocopherol complex solution (12 mg/ml) with 1 ml of reservoir solution (12% (w/v) PEG 6000,0.1 M sodium citrate (pH 5.6), 0.1 M LiSO4). Tetragonalcrystals were observed after two weeks and grew to anaverage size of 0.05 mm £ 0.05 mm £ 0.05 mm. Beforedata collection crystals were flash-cooled in a nitrogenstream at 110 K after raising the MPD concentrationof the crystallization solution to 25% (w/v). Data werecollected at the SLS synchrotron beamline X06SA (PSIVilligen) at 100 K, employing a MAR-165 CCD (MARX-ray-research, Hamburg, Germany) detector. Twopasses were made to collect high-resolution and over-loaded low-resolution reflections. The crystal-to-detectordistances varied between 160 mm and 180 mm forselenomethionine derivatives and between 120 mm and140 mm for the native and tocopherol complex data set.Oscillation angles varied between 1 and 28 per frame,exposure times ranged from one to three seconds. Datawere integrated and scaled with XDS42,43 (Table 1).

Structure solution and refinement

The structure was solved by SAD using SeMet-labeledprotein and the monoclinic crystal form which wasavailable first. Six out of eight expected seleniumpositions were determined using Shake-and-Bake ver-sion 2.1,44 phases were computed using SOLVE 2.0145

and RESOLVE.46 Further density modification, phaseextension and automatic model building was done byArpWarp47 using the native monoclinic data set. Refine-ment was effected using REFMAC,48 for model rebuild-ing the program O was used.49 Non-crystallographicsymmetry restraints were adjusted to give the lowestR-free value. Further details are given in Table 1.

The structure of the a-RRR-T complex was determinedby molecular replacement using CNS.50 Clear density forthe bound a-RRR-T was visible from the beginning.Refinement statistics are given in Table 1.

Protein Data Bank accession codes

Coordinates of both structures have been depositedin the RCSB Protein Data Bank with ID codes 1oipand 1oiz.

Acknowledgements

This work has been supported by the Universityof Berne, Switzerland, the Kontaktgruppe furForschungsfragen of the Basel industry and theBerner Hochschulstiftung. The help of all the staffat the Swiss Light Source in Villigen is highlyappreciated.

References

1. Burton, G. W., Joyce, A. & Ingold, K. U. (1983). Isvitamin E the only lipid-soluble, chain-breaking anti-oxidant in human blood plasma and erythrocytemembranes? Arch. Biochem. Biophys. 221, 281–290.

2. IUPAC-IUB Joint Commission on BiochemicalNomenclature (JCBN) (1982). Nomenclature of toco-pherols and related compounds. Recommendations1981. Eur. J. Biochem. 123, 473–475.

3. Burton, G. W., Traber, M. G., Acuff, R. V., Walters,D. N., Kayden, H., Hughes, L. & Ingold, K. U.(1998). Human plasma and tissue a-tocopherol con-centrations in response to supplementation withdeuterated natural and synthetic vitamin E. Am.J. Clin. Nutr. 67, 669–684.

4. Gallo-Torres, H. E. (1970). Obligatory role of bile forthe intestinal absorption of vitamin E. Lipids, 5,379–384.

5. Traber, M. G., Sokol, R. J., Burton, G. W., Ingold,K. U., Papas, A. M., Huffaker, J. E. & Kayden, H. J.(1990). Impaired ability of patients with familialisolated vitamin E deficiency to incorporate alpha-tocopherol into lipoproteins secreted by the liver.J. Clin. Invest. 85, 397–407.

6. Hosomi, A., Arita, M., Sato, Y., Kiyose, C., Ueda, T.,Igarashi, O. et al. (1997). Affinity for alpha-tocopheroltransfer protein as a determinant of the biologicalactivities of vitamin E analogs. FEBS Letters, 409,105–108.

7. Ingold, K. U., Burton, G. W., Foster, D. O., Hughes,L., Lindsay, D. A. & Webb, A. (1987). Biokinetics ofand discrimination between dietary RRR- and SRR-a-tocopherols in the male rat. Lipids, 22, 163–172.

8. Arita, M., Sato, Y., Miyata, A., Tanabe, T., Takahashi,E., Kayden, H. J. et al. (1995). Human alpha-toco-pherol transfer protein: cDNA cloning, expressionand chromosomal localization. Biochem. J. 306,437–443.

9. Kalikin, L. M., Bugeaud, E. M., Palmbos, P. L., Lyons,R. H., Jr & Petty, E. M. (2001). Genomic characteriz-ation of human SEC14L1 splice variants within a17q25 candidate tumor suppressor gene region andidentification of an unrelated embedded expressedsequence tag. Mamm. Genome, 12, 925–929.

10. Bankaitis, V. A., Aitken, J. R., Cleves, A. E. &Dowhan, W. (1990). An essential role for a phospho-lipid transfer protein in yeast Golgi function. Nature,347, 561–562.

11. Intres, R., Goldflam, S., Cook, J. R. & Crabb, J. W.(1994). Molecular cloning and structural analysis ofthe human gene encoding cellular retinaldehyde-binding protein. J. Biol. Chem. 269, 25411–25418.

12. Shibata, N., Arita, M., Misaki, Y., Dohmae, N., Takio,K., Ono, T. et al. (2001). Supernatant protein factor,which stimulates the conversion of squalene tolanosterol, is a cytosolic squalene transfer proteinand enhances cholesterol biosynthesis. Proc. NatlAcad. Sci. USA, 98, 2244–2249.

13. Aravind, L., Neuwald, A. F. & Ponting, C. P. (1999).Sec14p-like domains in NF1 and Dbl-like proteinsindicate lipid regulation of Ras and Rho signaling.Curr. Biol. 9, R195–R197.

14. Traber, M. G. & Arai, H. (1999). Molecular mechan-isms of vitamin E transport. Annu. Rev. Nutr. 19,343–355.

15. Ouahchi, K., Arita, M., Kayden, H., Hentati, F., BenHamida, M., Sokol, R. et al. (1995). Ataxia with iso-lated vitamin E deficiency is caused by mutations in

732 Structure of Human a-Tocopherol Transfer Protein

the alpha-tocopherol transfer protein. Nature Genet. 9,141–145.

16. Yokota, T., Igarashi, K., Uchihara, T., Jishage, K.,Tomita, H., Inaba, A. et al. (2001). Delayed-onsetataxia in mice lacking alpha-tocopherol transfer pro-tein: model for neuronal degeneration caused bychronic oxidative stress. Proc. Natl Acad. Sci. USA,98, 15185–15190.

17. Jishage, K., Arita, M., Igarashi, K., Iwata, T.,Watanabe, M., Ogawa, M. et al. (2001). Alpha-toco-pherol transfer protein is important for the normaldevelopment of placental labyrinthine trophoblastsin mice. J. Biol. Chem. 276, 1669–1672.

18. Zimmer, S., Stocker, A., Sarbolouki, M. N., Spycher,S. E., Sassoon, J. & Azzi, A. (2000). A novel humantocopherol-associated protein: cloning, in vitroexpression, and characterization. J. Biol. Chem. 275,25672–25680.

19. Stocker, A., Tomizaki, T., Schulze-Briese, C. &Baumann, U. (2002). Crystal structure of the humansupernatant protein factor. Structure (Camb.), 10,1533–1540.

20. Sha, B., Phillips, S. E., Bankaitis, V. A. & Luo, M.(1998). Crystal structure of the Saccharomycescerevisiae phosphatidylinositol-transfer protein.Nature, 391, 506–510.

21. Kleywegt, G. J. & Jones, T. A. (1994). Detection,delineation, measurement and display of cavities inmacromolecular structures. Acta Crystallog. sect. D,50, 178–185.

22. Lehmann, J., Martin, H. L., Lashley, E. L., Marshall,M. W. & Judd, J. T. (1986). Vitamin E in foods fromhigh and low linoleic acid diets. J. Am. Dietet. Assoc.86, 1208–1216.

23. Otzen, D. E., Rheinnecker, M. & Fersht, A. R. (1995).Structural factors contributing to the hydrophobiceffect: the partly exposed hydrophobic minicore inchymotrypsin inhibitor 2. Biochemistry, 34,13051–13058.

24. Traber, M. G., Burton, G. W., Ingold, K. U. & Kayden,H. J. (1990). RRR- and SRR-a-tocopherols aresecreted without discrimination in human chylo-microns, but RRR-a-tocopherol is preferentiallysecreted in very low density lipoproteins. J. LipidRes. 31, 675–685.

25. Traber, M. G., Rudel, L. L., Burton, G. W., Hughes, L.,Ingold, K. U. & Kayden, H. J. (1990). Nascent VLDLfrom liver perfusions of cynomolgus monkeys arepreferentially enriched in RRR- compared with SRR-a-tocopherol: studies using deuterated tocopherols.J. Lipid Res. 31, 687–694.

26. Weiser, H. & Vecchi, M. (1982). Stereoisomers ofa-tocopheryl acetate II. Biopotencies of all eightstereoisomers, individually or in mixtures, as deter-mined by rat resorption-gestation tests. Int. J. Vitam.Nutr. Res. 52, 351–370.

27. Weiser, H., Riss, G. & Kormann, A. W. (1996). Bio-discrimination of the eight a-tocopherol stereo-isomers results in preferential accumulation of thefour 2R forms in tissues and plasma of rats. J. Nutr.126, 2539–2549.

28. Schouten, A., Agianian, B., Westerman, J., Kroon, J.,Wirtz, K. W. & Gros, P. (2002). Structure of apo-phos-phatidylinositol transfer protein a provides insightinto membrane association. EMBO J. 21, 2117–2121.

29. Yoder, M. D., Thomas, L. M., Tremblay, J. M., Oliver,R. L., Yarbrough, L. R. & Helmkamp, G. M., Jr(2001). Structure of a multifunctional protein.Mammalian phosphatidylinositol transfer protein

complexed with phosphatidylcholine. J. Biol. Chem.276, 9246–9252.

30. van Tiel, C. M., Schouten, A., Snoek, G. T., Gros, P. &Wirtz, K. W. (2002). The structure of phosphatidyl-inositol transfer protein a reveals sites for phospho-lipid binding and membrane association with majorimplications for its function. FEBS Letters, 531, 69–73.

31. Yokota, T., Shiojiri, T., Gotoda, T. & Arai, H. (1996).Retinitis pigmentosa and ataxia caused by amutation in the gene for the a-tocopherol-transferprotein. N. Engl. J. Med. 335, 1770–1771.

32. Yokota, T., Uchihara, T., Kumagai, J., Shiojiri, T.,Pang, J. J., Arita, M. et al. (2000). Postmortem studyof ataxia with retinitis pigmentosa by mutation ofthe alpha-tocopherol transfer protein gene. J. Neurol.Neurosurg. Psychiatry, 68, 521–525.

33. Harding, A. E. (1993). Clinical features and classifi-cation of inherited ataxias. Advan. Neurol. 61, 1–14.

34. Cavalier, L., Ouahchi, K., Kayden, H. J., Di Donato,S., Reutenauer, S., Mande, J. L. & Koenig, M. (1998).Ataxia with isolated vitamin E deficiency: hetero-geneity of mutations and phenotypic variability in alarge number of families. Am. J. Hum. Genet. 62,301–310.

35. Gotoda, T., Arita, M., Arai, H., Inoue, K., Yokota, T.,Fukuo, Y. et al. (1995). Adult-onset spinocerebellardysfunction caused by a mutation in the gene forthe alpha-tocopherol-transfer protein (see com-ments). N. Engl. J. Med. 333, 1313–1318.

36. Benomar, A., Yahyaoui, M., Meggouh, F., Bouhouche,A., Boutchich, M., Bouslam, N. et al. (2002). Clinicalcomparison between AVED patients with 744 del Amutation and Friedreich ataxia with GAA expansionin 15 Moroccan families. J. Neurol. Sci. 198, 25–29.

37. Yokota, T., Shiojiri, T., Gotoda, T., Arita, M., Arai, H.,Ohga, T. et al. (1997). Friedreich-like ataxia withretinitis pigmentosa caused by the His101Glnmutation of the alpha-tocopherol transfer proteingene. Ann. Neurol. 41, 826–832.

38. Phillips, S. E., Sha, B., Topalof, L., Xie, Z., Alb, J. G.,Klenchin, V. A. et al. (1999). Yeast Sec14p deficient inphosphatidylinositol transfer activity is functional invivo. Mol. Cell, 4, 187–197.

39. Goa, J. (1953). A Micro Biuret method for proteindetermination. Determination of total protein in cere-brospinal fluid. Scand. J. Clin. Lab. Invest. 5, 218–222.

40. Van Duyne, G. D., Standaert, R. F., Karplus, P. A.,Schreiber, S. L. & Clardy, J. (1993). Atomic structuresof the human immunophilin FKBP-12 complexeswith FK506 and rapamycin. J. Mol. Biol. 229,105–124.

41. Panagabko, C., Morley, S., Neely, S., Lei, H., Manor,D. & Atkinson, J. (2002). Expression and refolding ofrecombinant human alpha-tocopherol transferprotein capable of specific a-tocopherol binding.Protein Expr. Purif. 24, 395–403.

42. Kabsch, W. (1993). Automatic processing of rotationdiffraction data from crystals of initially unknownsymmetry and cell constants. J. Appl. Crystallog. 26,795–800.

43. Kabsch, W. (1988). Evaluation of single crystal X-raydiffraction data from a position sensitive detector.J. Appl. Crystallog. 21, 916–924.

44. Hauptman, H. A. (1997). Shake-and-bake: analgorithm for automatic solution ab initio of crystalstructures. Methods Enzymol. 277, 3–13.

45. Terwilliger, T. C. & Berendzen, J. (1999). AutomatedMAD and MIR structure solution. Acta Crystallog.sect. D, 55, 849–861.

Structure of Human a-Tocopherol Transfer Protein 733

46. Terwilliger, T. C. & Berendzen, J. (1999). Discrimi-nation of solvent from protein regions in nativeFouriers as a means of evaluating heavy-atomsolutions in the MIR and MAD methods. ActaCrystallog. sect. D, 55, 501–505.

47. Perrakis, A., Morris, R. & Lamzin, V. S. (1999). Auto-mated protein model building combined withiterative structure refinement. Nature Struct. Biol. 6,458–463.

48. Murshudov, G. N., Vagin, A. A. & Dodson, E. J.(1997). Refinement of macromolecular structures bythe maximum likelihood method. Acta Crystallog.sect. D, 53, 240–255.

49. Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard

(1991). Improved methods for building proteinmodels in electron density maps and the location oferrors in these models. Acta Crystallog. sect. A, 47,110–119.

50. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano,W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998).Crystallography and NMR system: a new softwaresuite for macromolecular structure determination.Acta Crystallog. sect. D, 54, 905–921.

51. Wallace, A. C., Laskowski, R. A. & Thornton, J. M.(1995). LIGPLOT: a program to generate schematicdiagrams of protein-ligand interactions. Protein Eng.8, 127–134.

Edited by R. Huber

(Received 26 March 2003; received in revised form 28 May 2003; accepted 30 May 2003)

734 Structure of Human a-Tocopherol Transfer Protein

View publication statsView publication stats